#hashtag
# AI Agents
Agents are an advanced AI systems designed to make decisions and take actions like sequential reasoning, planning, execution, and memory. Acts on its own step by step without human intervention. LLMs are the AI agents which completes automatic tasks with a well-defined goals and objectives.
Agents perform data analysis, strategic planning, data retrieval, and the ability to learn from past actions to solve complex problems in various business use cases. To complete various tasks which LLM agent requires a structured plan, a reliable memory to track agent functioning, and necessary tools. LLM agent’s workflow components are agent, planning, memory, and tools.
## 1. Memory
The memory of agents supports complex tasks what previously completed.
The memory of LLM agents is composed of two main memory types. Short-term memory captures important information related conversation, ongoing discussion, supports the model response to the immediate context, short term memory is temporary. Long-term memory captures and storing important information from previous interactions. Understanding patterns, learning from previous tasks, and retrieval the information to make better decisions in future interactions.
## 2. Planning
Agents can make reasoning, split down complex tasks into smaller tasks, and develop specific plans. As tasks evolve agents can also reflect on and tune their plans with respect to real-time problems. Planning typically involves two main components one is plan formulation and another one plan reflection.
## 3. Tools
Tools in agents connect with external sources which to perform certain tasks extracting information from databases and querying. It follows specific workflows to carry out tasks, gather observations, information needed to complete subtasks and fulfil user requests.
--MATHEMATICS GRADUATE
2moI agree @